This is a single institutional study involving retrospective analysis of prospectively collected data. The following pathological criteria were used for patient selection for neoadjuvant systemic treatment: (1) LABC (not including node positive T1 and T2 LABC cases), (2) LOBC, (3) luminal HER2-positive and HER-positive (non-luminal) tumors greater than 2 cm; (4) Triple Negative Breast Cancer (TNBC) greater than 3 cm, (5) luminal HER2 negative large T2/T3 tumors, and (6) any patient with clinically or histologically proven positive axillary nodes. Patients with demonstratable metastatic disease were excluded. All patients underwent definitive breast surgery post-NACT in the period between January 2015 and December 2020. Patients were deemed qualified for upfront mastectomy based on the following criteria: (1) LABC; (2) LOBC; that is, Stage IIB and above (T3N0 but not T2N1) as described by Simos et al. (29); (3) small breast volume to tumor ratio; and (4) multicentric/multifocal tumors. Having upfront qualification for mastectomy enabled us to compute conversion rates to BCS/OBS post-NACT. The plan of management (NACT and proposed surgical plan) was discussed with the patient, and written informed consent was obtained for each step/procedure, for each patient.

Data included demography, medical history, clinical findings, pathological reports (diagnostic biopsy and surgical histopathology including immunohistochemistry), details of NACT, surgical intervention, pre- and post-operative images of patients, post-surgery complications, follow-up details, and PROMs.

Triple assessment based on clinical examination, imaging and image-guided core needle biopsy was routinely used to establish a diagnosis. Systemic staging was assessed based on Positron Emission Tomography (PET) imaging. Patients were selected for neoadjuvant treatment based on decisions made by the multidisciplinary team (MDT). After clinical staging, NACT was administered based on NCCN guidelines.

After every other cycle of chemotherapy and at the completion of NACT, the patient was monitored clinically and radiologically. Clinical response of the primary tumor to therapy was calculated according to the RECIST Ver 1.1 criteria (30). The pathological response was determined by comparing the pre-therapy clinical stage with the stage at final histopath.

For all cases, the residual tumor was pre-operatively localized by ultrasound (USG) imaging. In cases where the tumor responded to NACT, the residual tumor was clipped mid-NACT at ~1cm size by USG guided insertion of liga clips. At least four clips were used to facilitate identification by intra-operative USG ( Figure 1 ). We believe that accurate delineation of the center is difficult in a 4- to 5-cm tumor and therefore needs to be done after response to therapy and reduction in tumor volume. Mid-NACT localization after good NACT response mitigated the requirement of another localization procedure such as wire guidance before the definitive surgery. Pre-operative localization by mammography-guided wire bracketing was done only where there was extensive residual calcification post-NACT. Intra-operative USG was used to target the clipped center, and a wide excision of the clip-bearing area was carried out. Intra-operative specimen mammography and USG were used to confirm that all wire-bracketed calcifications and/or clipped specimen were excised with negative margins. Figure 1 also shows specimen mammogram from clipped residual tumors that were T0 and T1 on final pathology.

In this report, paraffin slides of margin tissue were accessed from the pathology laboratory to verify the negative margin outcome. Slides of tissue assessed at frozen were quality checked for tissue integrity on Mantra (PerkinElmer). Frozen and pathology slides were imaged on an OS-15 (OptraSCAN, San Jose, CA, USA) bright field digital scanner. These digital images were blinded and shared from our server with the second pathologist (SB) to assess concordance in margin assessment.

In every case, wide local excision of the residual tumor was identified by palpation as well as USG of the residual tumor and the marker clips used for localization. Post-excision margin adequacy was assessed on the table by USG or by specimen mammography. In addition, all specimens were sent for frozen section margin assessment to the pathologist. Any close margin (1–2 mm) seen on specimen mammogram or reported on frozen section (including focally positive margins) was revised in the same sitting before restoration of the breast form. Margin guidelines based on the consensus on margin safety by the American Society of Clinical Oncology, Society of Surgical Oncology, and American Society for Radiation Oncology Consensus Guideline (31) were followed.

The surgical plan was determined based on assessment at diagnosis, post-therapy clinical-radiological assessment of residual disease, residual tumor location, and extent of calcification, breast size, and ptosis. Patients were counseled for safety and esthetic outcomes. Final decisions were based on patient choice. Volume displacement and volume replacement techniques were employed in all cases where conventional breast conservation was deemed unsuitable. In cases where breast conservation was not feasible mastectomy with immediate whole breast reconstruction was performed. Very few patients did not undergo reconstruction. The surgical procedures used are described here.

BCS with primary breast closure was performed in patients with small residual tumors with adequately sized breasts or tumors in favorable locations.

Oncoplastic surgeries have been classified based on the recent recommendations of the American Society of Breast Surgeons (32) and as expanded on by Silverstein (20). As defined in these recommendations, oncoplastic surgery refers purely to oncoplastic breast conservation surgery. Oncoplastic techniques performed were simple: Level 1, complex; Level 2, volume displacement and more complex; and Level 3, volume replacement and by perforator flap in most cases. Mini-Latissimus dorsi (LD) flaps were only used in surgeries performed prior to 2019 before perforator flaps were incorporated in routine practice.

In certain situations, a complete mastectomy was performed. Seventy percent cases of mastectomy were followed by immediate breast reconstruction by implant or LD flaps. Dermal sling was employed for larger ptotic breasts, which has been shown to be a safe procedure (33). For cases with small breasts, an Advanced–Lower Dermal Sling (A-LDS) was used in place of an acellular dermal matrix (34).

Post-surgical oncological management decisions regarding chemoradiation protocols were undertaken by a multidisciplinary clinical team in accordance with guidelines.

Surgical outcome data were recorded as minor or major based on the presence or absence of surgical complications such as seroma, delayed wound healing, fat necrosis, lymphedema, and infection and the interventions required to treat them. Major complications were those that required surgical intervention, whereas minor complications could be treated by therapy in the clinic.

Margin status was assessed by histopathology of frozen sections intra-operatively and paraffin section post-operatively. Patients were carefully monitored for local and distant recurrences, quarterly for the first 2 years post-surgery and then every 6 months. Suspicious symptoms or signs were assessed by appropriate imaging and histological confirmation wherever feasible.

Patients were evaluated for patient reported outcome measures (PROMs) using the BREAST-Q questionnaire (35) at 1-year post-surgery follow-up. BREAST-Q questionnaire was offered to all patients at the completion of 1-year post-surgery follow-up, and data are presented for all patients who responded.

Data were analyzed using Fishers Exact Count Statistic (when less than five cases), chi-square test, Wilcoxon Rank Test, and Student’s t-test using the stat and BSDA (36) package in R Ver 4.0 (37). Median follow-up was calculated using the reverse KM method of Schemper and Smith (38) using the Prodlim (39) and Hmisc (40) packages in R Ver 4.0. All graphs were plotted using ggpubR (41) in R Ver 4.0. Survival parameters were calculated using the survival (42) package in R. Kaplan–Meier plots were plotted using the survminer (43) package.

One hundred eighty breast cancer patients treated with neoadjuvant therapy and operated between January 2015 and December 2020 were included in the study. Table 1 describes the demographic and clinical features of the patients included in the cohort. The median age of the patients was 50 years (range: 23–75). Most cases (41%, 74 of 180) were luminal HER2 negative, followed by TNBC (31%, 57 of 180) and HER2 + 27% (49 of 180). HER2-positive cases comprise 21 luminal HER2-positive cases and 28 HER2-positive non-luminal cases.

A majority of cases (86%) were diagnosed at biopsy as IDC (155 of 180), and the rest were IDC with DCIS (9 of 180, 5%) and ILC (7 of 180, 4%). Most tumors were T2 at presentation (63%, 113 of 180), and 85% of patients were node positive (153 of 180). Most of these cases were N2 on imaging (56%, 86 of 153). A significant number of cases were clinically stage IIIA at presentation (38%, 69 of 180). The three cases that were treated with neoadjuvant therapy post an excisional biopsy performed at other centers. Excision biopsy is performed on lumpectomy specimens. These are the results of surgical procedures performed for diagnosis, immediately after a suspicious imaging report or clinical examination, without any initial biopsy. This is a common practice in many centers. Such patients often come to our center for reconstructive surgery to regain breast shape, remove scars, and so forth. This is a common practice in many centers. Such patients often come to our center for reconstructive surgery to regain breast shape, remove scars, and so forth. Since margin positivity is often not assessed after such procedures, patients are always assessed for residual tumor by imaging or a histopathological review of specimen tissue when available. The three cases included here came to us post such a procedure from other centers with reported positive margins. Two of these cases also had positive nodes that had not been treated. The cases included in this cohort were young age (< 40 years) and offered NACT prior to surgery.

Among patients with luminal HER2-negative tumors, 49% (36 of 74) received hormonal therapy and 51% (38 of 76) were given Anthracycline- or Anthracycline- and Paclitaxel-based chemotherapy. Patients on hormone therapy were significantly older (mean age: 61 + 9.6, (range: 41–75 years) than the chemotherapy group (mean age: 48 + 10.3, (range: 29–67 years), p = 6.39e−07), and all except 1 were post-menopausal. Seventy-seven percent (38 of 49) of HER2-positive (luminal and non-luminal) patients received trastuzumab, a very high rate in an Indian setting for the use of HER2-directed therapy due to the prohibitive cost involved (44). Fifty-eight percent of TNBC patients (33 of 56) received AC followed by a Taxane, and the rest were treated with protocols such as AC alone, Fluorouracil, Epirubicin, Cyclophosphamide (FEC) with Taxane and Taxane with platinum drug as per guidelines. These variations in drug regimen are unavoidable confounders in our data and are a result of the socio-economic realities that patients face during treatment. In some cases where the delay in surgery and/or increased costs due to neoadjuvant therapy proved unacceptable, the preferred regiment of AC + taxane was cut down to four cycles of AC alone.

OBS dominated as the choice of surgery in our cohort ( Figure 2 ). Seventeen of 20 patients who had complete clinical response underwent OBS, while 53 of 73 patients with partial response were treated with OBS. Patients with progressive disease had the highest percent of mastectomy with immediate breast reconstruction (31%) or without reconstruction (9%).

The clinical response to treatment was determined by RECIST criteria Ver 1.1. (30). The pattern of response based on molecular subtype is similar to the reported data (45). Complete response (pathological) was high in TNBC and HER2-positive non-luminal (pCR in 48 and 60%, respectively), and low in luminal HER2-positive and luminal HER2-negative (pCR in 29 and 10% cases, respectively). The clinical and pathological response is shown in Figure 3 .

Patients were classified as qualified for upfront BCS (99 of 180) or upfront mastectomy (81 of 180). The basis for qualification for mastectomy is summarized in Table 2 . Definitive surgery was performed at the end of the prescribed NACT period, based on the clinical response of the tumor to NACT ( Figure 2 ).

Out of 81 cases that initially qualified for mastectomy, 59 showed a good response to NACT and hence were converted to BCS/OBS with the conversion rate from mastectomy to BCS/OBS of 72.8%. The rest underwent mastectomy with reconstruction (15) or mastectomy alone (7) (see Table 3 and Supplementary Table S1 ). These patients received mastectomy despite being pre-therapy candidates for breast conservation. Two of these patients had progressed on treatment, and two patients had a complete response but missed mid-NACT USG examination for clip placement and therefore had to be treated with mastectomy. For the rest of the cases, the surgery choice was based on patient preference. In some of these cases, after inadequate response to therapy and with worrisome biology, the patients were not convinced of the safety of breast conservation and were therefore treated with a mastectomy.

Overall, 142 (80%) breast-conserving surgeries were performed in the cohort. A substantial proportion (79%) of the breast conservative operations is oncoplastic procedures (OBS) (121 of 142). The types of OBS carried out in the cohort and for the subset that qualified for upfront mastectomy are shown in Table 4 . The frequency and type of oncoplastic surgeries in the mastectomy qualified cohort (81 of 180) and those amenable to upfront BCS (99 of 180) are not significantly different. The complexity of oncoplastic surgeries varied from volume displacement techniques (Level 1 techniques) such as rotational mammoplasty to more complex skilled procedures for extreme oncoplasty such as therapeutic mammoplasty and perforator flaps. Volume replacement procedures (46 of 142) following partial mastectomy utilized perforator flaps in most cases (83%, 38 of 46) and for the rest mini-LD (17%, 7 of 46) was performed, if such procedures could achieve superior outcomes. Mini-LD partial reconstructions were mostly done before 2019 when we had yet to adopt perforator flap as a routine technique in our practice. Examples of imaging findings and oncoplastic procedures are illustrated in Figure 4 .

Patients who were selected for volume displacement Level 2 surgeries were also counseled for contralateral symmetrization. Only patients who consented for symmetrization were treated with this oncoplastic technique (40 cases). In addition, three cases of perforator flap volume replacement and 11 cases of immediate breast reconstruction received simultaneous symmetrization.

The tissue volume after resection was significantly smaller in BCS (82 ± 53 cc) as compared with average resection volume in OBS procedures (235 ± 280 cc, p = 0.0012) ( Figure 5A ). To ensure the oncological safety of breast conservation, larger volumes of excisions were required in some cases; hence, we had to adopt OBS techniques. The volume and the location of the excision required dictated the complexity of the OBS technique required ( Figure 5A ).

A total of 40% patients (72 of 180) received adjuvant chemotherapy. A majority of HER2-positive (luminal and non-luminal) patients received adjuvant trastuzumab (57%), while only 45% (25 of 56) TNBC patients and 49% (36 of 74) luminal HER2-negative patients received adjuvant chemo- or hormonal therapy. Most patients received adjuvant radiotherapy (RT) (92%—167 of 180). Of the patients who did not receive RT, seven were lost to follow up, and six patients refused RT. Of these six patients, none had recurred at the time of last follow-up (average follow-up for these patients = 39 months).

The frequency and type of post-surgical complications observed in the NACT cohort in the first year post-surgery are shown in Table 5 . Minor complications including seroma, fat necrosis, and insignificant delayed wound healing in OBS surgeries were observed in 19% of cases (23 of 121) and in 16% of immediate breast reconstruction surgeries (4 of 25). Two patients had major complications due to delayed wound healing post and OBS surgery, leading to flap necrosis and completion mastectomy in one case and abdominal flap graft in the other. Delayed wound healing in one case of breast reconstruction led to loss of implant (more than 1 year post-surgery). All other complications were minor and were treated with either no intervention or conservatively without needing to go back to the theater.

The median follow-up for this cohort is 38 months (minimum 6.1 months and maximum 86 months, i.e., 7.2 years). Sixteen cases (9%) have less than 1-year follow-up. At median follow-up (38 months), the entire cohort had an overall survival (OS) rate of 90.5%, distant disease–free survival of 84%, and loco-regional recurrence free survival at 93% ( Figure 6 upper panel and Table 6 ). The 142 breast conservation–treated cases had an OS rate of 91.3%, distant disease free at 86.0%, and LRR-free survival of 93.5%. ( Figure 6 lower panel). These are acceptable rates for post-NACT–treated cohorts. The Kaplan–Meier plot for surgery stratified disease free and OS is shown in Figure 6 . Surgery-type specific local and distant recurrences that occurred during median follow-up time are shown in Table 7 . Further details for each breast conservation type are shown in Supplementary Table S2 . Cox proportional hazards showed no significant difference at median follow-up (38 months) between breast conservation (conventional BCS and OBS) and mastectomy (mastectomy with or without immediate breast reconstruction) for both distant and local disease free and OS (data not shown). This pattern was seen even in the comparison of the outcomes for the mastectomy qualified versus unqualified cohort ( Figure 7 ). Subtype-stratified distant and local disease free and OS was significantly worse for TNBC subtype. Multivariate cox proportional hazards showed that only TNBC subtype and pathological node status are significantly associated with disease free outcomes in this dataset (data not shown).

The PROM scores from the BREAST-Q questionnaire (35) (Questions 1–5) are shown in Table 8 . The data represent 1-year PROM scores collected for 18 of 25 patients who had a mastectomy with immediate breast reconstruction, and 72 of 121 patients who underwent OBS (22 volume displacement Level 1, 25 volume displacement Level 2, and 25 volume replacement—perforator flaps). All parameters are satisfactory in all three surgery types, that is, breast reconstruction, BCS, and OBS (values range from 67% to 88%). The variation among surgery types for scores of Question 1 (satisfaction with breasts) was significantly different in ANOVA analysis. OBS showed significantly better satisfaction scores compared with reconstruction, for Question 1 ( Table 8 ) (P < 0.001) by Mann–Whitney Test.